I believe that matching computer to worktype is a planned feature in the v5 server. The details are still being worked out I think.

[QUOTE=garo;124749]I believe that matching computer to worktype is a planned feature in the v5 server. The details are still being worked out I think.[/QUOTE]

The default worktype for a fast machine like a core 2 duo will be LL testing - not trial factoring.

Whereas a v4 client always made the decision of the type of work to get, the v5 server uses PHP code to make this decision. I can take into account several variables in deciding what to assign. After a few months of v5 operation we should have a good idea on ways to improve the decision making process. In general though I believe in obeying the user's work preference when he has overridden the default "do what makes the most sense".

Also note, that the server is not handing out the harder 67->68 and 68->69 TF assignments now. This is because no beta testers are doing P-1 factoring. These harder TFs are only assigned after P-1 factoring is done.

[QUOTE=Prime95;124787]Also note, that the server is not handing out the harder 67->68 and 68->69 TF assignments now. This is because no beta testers are doing P-1 factoring. These harder TFs are only assigned after P-1 factoring is done.[/QUOTE]Is that still by design? Would you like me (or anyone) to try some P-1?

[QUOTE=James Heinrich;124813]Is that still by design? Would you like me (or anyone) to try some P-1?[/QUOTE]

Feel free to sign up for the "P-1 on large Mersennes" (or whatever its called) work preference. There is no reason why it shouldn't work - I tried it on a few exponents a couple months ago.

[QUOTE=Prime95;124818]Feel free to sign up for the "P-1 on large Mersennes" (or whatever its called) work preference. There is no reason why it shouldn't work - I tried it on a few exponents a couple months ago.[/QUOTE]It's just called "P-1 factoring". I set one worker to work on that. Should be done first exponent ~06Feb.

I've done a couple of dozen P-1's I have client working on them. Slow work though avg 2.03 days

Thought I would mention also that since switching to 25.6 I have not noticed any problems so far with the client. The server pages do odd things from time to time but otherwise things seem to be going ok.

ECM-F type of work
 

Would it be possible to assign just one ECM curve at time for small Fermat numbers?

Luigi

[QUOTE=ET_;124862]Would it be possible to assign just one ECM curve at time for small Fermat numbers?

Luigi[/QUOTE]

I think this is exactly what Prime95 does: you get one ECM curve assigned for a certain Fermat number, or you may get some assignments of one ECM curve each.

[QUOTE=Andi47;124874]I think this is exactly what Prime95 does: you get one ECM curve assigned for a certain Fermat number, or you may get some assignments of one ECM curve each.[/QUOTE]

I got three assignments of F21, for 3 curves EACH :-)

Maybe it worked that way because we were short of TF work, and now the server behaves differently. :ermm:

It would be nice to know how it works now.

Luigi

Right now it always assigns three curves for ECM work. I can change that based on exponent, B1 values, and cpu speed.